Methods and apparatus to trigger calibration of a sensor node using machine learning

Abstract

Methods, apparatus, systems and articles of manufacture to trigger calibration of a sensor node using machine learning are disclosed. An example apparatus includes a machine learning model trainer to train a machine learning model using first sensor data collected from a sensor node. A disturbance forecaster is to, using the machine learning model and second sensor data, forecast a temporal disturbance to a communication of the sensor node. A communications processor is to transmit a first calibration trigger in response to a determination that a start of the temporal disturbance is forecasted and a determination that a first calibration trigger has not been sent.

What is claimed is: 
     
       1. An apparatus to trigger calibration of sensor nodes, the apparatus comprising:
 a machine learning model trainer to train a machine learning model using first sensor data collected from a sensor node; 
 a disturbance forecaster to, using the machine learning model and second sensor data, forecast a temporal disturbance including an interference to a communication ability of the sensor node; and 
 a communications processor to:
 transmit a first calibration trigger in response to a first determination that a start of the temporal disturbance is forecasted and a second determination that the first calibration trigger has not been sent; and 
 transmit a second calibration trigger in response to a third determination that an end of the temporal disturbance is forecasted and a fourth determination that the second calibration trigger has not been sent. 
 
 
     
     
       2. The apparatus of  claim 1 , wherein the machine learning model trainer is further to determine that the first sensor data is sufficient for training the machine learning model, and perform the training of the machine learning model in response to determining that the first sensor data is sufficient. 
     
     
       3. The apparatus of  claim 1 , wherein the machine learning model trainer is further to re-train the machine learning model using the second sensor data. 
     
     
       4. The apparatus of  claim 1 , wherein the communications processor is further to, in response to receipt of a communication from the sensor node, extract the second sensor data from the communication, determine whether the second sensor data is already stored in a data store, and in response to determining that the second sensor data is already stored in the data store, not store the second sensor data in the data store. 
     
     
       5. The apparatus of  claim 1 , further including a beacon counter, and the communications processor is further to, in response to receipt of a first communication indicating that the sensor node is to begin a calibration sequence, initialize a value of the beacon counter. 
     
     
       6. The apparatus of  claim 5 , wherein the communications processor is to, in response to receipt of a second communication, increment the value of the beacon counter. 
     
     
       7. The apparatus of  claim 6 , wherein the communications processor is to, in response to receipt of a third communication, transmit the value of the beacon counter to the sensor node. 
     
     
       8. At least one non-transitory machine-readable storage medium comprising instructions that, when executed, cause at least one processor to at least:
 train a machine learning model in response to determining that first sensor data collected from a sensor node is sufficient for training the machine learning model; 
 access second sensor data collected from the sensor node; 
 use the machine learning model and the second sensor data to forecast a temporal disturbance including an interference to a communication ability of the sensor node; 
 in response to a first determination that a start of the temporal disturbance is forecasted and a second determination that a first calibration trigger has not been sent, transmit the first calibration trigger; and 
 in response to a third determination that an end of the temporal disturbance is forecasted and a fourth determination that a second calibration trigger has not been sent, transmit the second calibration trigger. 
 
     
     
       9. The at least one non-transitory machine-readable storage medium of  claim 8 , wherein the instructions, when executed, further cause the at least one processor to determine that the first sensor data is sufficient for the training of the machine learning model when the first sensor data represents data collected over a threshold amount of time. 
     
     
       10. The at least one non-transitory machine-readable storage medium of  claim 8 , wherein the instructions, when executed, further cause the at least one processor to re-train the machine learning model using the second sensor data. 
     
     
       11. The at least one non-transitory machine-readable storage medium of  claim 8 , wherein the instructions, when executed, further cause the at least one processor to, in response to receipt of a communication from the sensor node:
 extract the second sensor data from the communication; 
 determine whether the second sensor data is already stored in a data store; and 
 in response to determining that the second sensor data is already stored in the data stored, not store the second sensor data in the data store. 
 
     
     
       12. The at least one non-transitory machine-readable storage medium of  claim 8 , wherein the instructions, when executed, further cause the at least one processor to, in response to receipt of a first communication indicating that the sensor node is to begin a calibration sequence, initialize a value of a beacon counter. 
     
     
       13. The at least one non-transitory machine-readable storage medium of  claim 12 , wherein the instructions, when executed, further cause the at least one processor to, in response to receipt of a second communication, increment the value of the beacon counter. 
     
     
       14. The at least one non-transitory machine-readable storage medium of  claim 13 , wherein the second communication is a beacon frame. 
     
     
       15. The at least one non-transitory machine-readable storage medium of  claim 13 , wherein the instructions, when executed, further cause the at least one processor to, in response to receipt of a third communication, transmit the value of the beacon counter to the sensor node. 
     
     
       16. An apparatus to trigger calibration of sensor nodes, the apparatus comprising:
 means for training a machine learning model using first sensor data collected from a sensor node; 
 means for forecasting, using the machine learning model and second sensor data, a temporal disturbance including an interference to a communication ability of the sensor node; and 
 means for transmitting to:
 transmit a first calibration trigger in response to a first determination that a start of the temporal disturbance is forecasted and a second determination that the first calibration trigger has not been sent; and 
 transmit a second calibration trigger in response to a third determination that an end of the temporal disturbance is forecasted and a fourth determination that the second calibration trigger has not been sent. 
 
 
     
     
       17. The apparatus of  claim 16 , wherein the means for training is to determine that the first sensor data is sufficient for the training of the machine learning model, the means for training to perform the training of the machine learning model in response to determining that the first sensor data is sufficient. 
     
     
       18. The apparatus of  claim 16 , wherein the means for training is to re-train the machine learning model using the second sensor data. 
     
     
       19. The apparatus of  claim 16 , wherein the means for transmitting is further to, in response to receipt of a communication from the sensor node, extract the second sensor data from the communication, determine whether the second sensor data is already stored in a data store, and in response to determining that the second sensor data is already stored in the data stored, not store the second sensor data in the data store.